Classification of antiarrhythmic agents

This chapter is (barely) relevant to Section G8(iv)  of the 2017 CICM Primary Syllabus, which asks the exam candidate to "understand the pharmacology of antiarrhythmic drugs".  These are a favourite of the college. For a variety of fairly commonsense reasons, there is an emphasis on testing the trainee's understanding of these drugs, as they are at the same time common, powerful, dangerous, and often poorly understood.  However, historical CICM examiner focus has been interestingly narrow: 

• Question 14 from the first paper of 2019 (digoxin vs. sotalol)
• Question 2 from the second paper of 2018 (amiodarone vs digoxin)
• Question 21 from the second paper of 2016 (classification with a focus on sotalol)
• Question 11 from the first paper of 2016 (amiodarone side effects)
• Question 13 from the second paper of 2014 (amiodarone)
• Question 9 from the second paper of 2012 (classification with a focus on Class 1 agents)
• Question 22 from the second paper of 2010 (amiodarone vs digoxin)
• Question 5 from the second paper of 2009 (digoxin)
• Question 5 from the second paper of 2008 (amiodarone)

Do you notice a trend here? There is clearly some thirst among the examiners for a) certain iodinated compounds and b) cardiac glycosides. Owing to this clearly preferential treatment, each of these drugs has been given a chapter of its own, all the better to be familiar with their properties. The other agents have been piled unceremoniously into the wastebasket chapter below, mainly because the topic of antiarrhythmic classification has only ever appeared in one SAQ.  What follows is a merciless oversimplification of a fascinating topic, but even this is probably excessive, as the college clearly just wanted people to earn three marks by dumbly listing the Vaughan Williams classes.

Oh well:

Realistically, the revising exam candiate would find that the Part One entry contains all the essential information required to pass these SAQs. It also helps that it is presented in a highly condensed form, free from verbal greebles and nurnies. For something published in a respectable journal (i.e. one which you have to pay to get published in), one could use something like Capucci et al (1998). This probably represents some sort of maximum of what a normal exam candidate should be expected to absorb. On the other hand, for the reader with an infinite capacity for minute pharmacological detail, the 1998 report by Carmeliet & Mubagwa will satisfy even the hungriest brain parasites. It is 72 pages long, and dense like a neutron star. 

Classification systems for antiarrhythmic agents

For the purposes of studying for the CICM First part exam, there is really only one system. One might occasionally hear it being referred to as the "Vaughan and Williams" or "Vaughan-Williams" classification, which is, of course, inaccurate because it is named after Miles Vaughan Williams, the celebrated pharmacologist and ninety-year-old fitness guru. His classification system was first presented in April of 1970, at the Symposium on Cardiac Arrhythmias in Elsinore, Denmark (yes, that Elsinore). The original 826-page symposium is not available electronically, and the nearest physical copy lays 110km north of the author, at the University of Newcastle Auschmuty Library. Fortunately, five years later, it was incorporated into a textbook, through which archaeologists can discern its original shape. 

• Class I: drugs which "interfere directly with depolarization", eg. quinidine
• Class II: drugs with "antisympathetic action", eg. beta-blockers
• Class III:  drugs which "prolong the duration of the action potential", eg amiodarone

It seems like a possible fourth class was also being considered in 1975, as verapamil had recently appeared onto the scene and had clearly demonstrated antiarrhythmic properties. However, it was not formally added at this stage.  In fact, like everything that gets incorporated into a textbook, this three-class system turned out to be extremely tenacious and remained basically unchanged well into the 1980s even as novel agents were developed and new physiology research had come to light.  The next reassessment of the Vaughan Williams classification was his own (1984), where he rephrased the class definitions, added calcium channel blockers, and further subdivided Class I into Ia, Ib and Ic on clinical grounds:

• Class I: fast sodium channel blockers:
  • Class Ia: prolong the action potential (eg. quinidine)
  • Class Ib: shortens the action potential (eg. lignocaine)
  • Class Ic: no effect on the action potential (eg. flecainide)
• Class II: Beta blockers (eg. metoprolol)
• Class III: Potassium channel blockers (eg. sotalol and amiodarone)
• Class IV: calcium channel blockers (eg. verapamil and diltiazem)

However, the modern reader will readily point out that not only do we have multiple other agents currently used to treat or prevent arrhythmias, but there are also many agents which were already well established in he 1980s (eg. digoxin) which seem to be unfairly omitted from this classification system. Moreover, many among the listed drugs have multiple effects across multiple classes (famously, amiodarone).  In response to these shortcomings, Rosen & Schwartz (1991) offered to reclassify antiarrhythmic drugs without trying to pigeonhole them into limiting categories. They called their system "The Sicilian Gambit",  in reference to an opening chess move which involves the sacrifice of pieces in order to achieve a strategic advantage. Confusingly, they called it "Sicilian" because the European Societ of Cardiology met in Sicily that year, and not because of the Sicilian defence which is another popular opening move.  These random details notwithstanding, the authors had a rather noble purpose: to bring the classification of antiarrhythmics out of its Vaughn Williams disarray, and to present the agents in a spreadsheet which illustrates their multiple simultaneous effects, like so:

Unfortunately, judging by the limited acceptance of this schema, it appears that everybody else had preferred disarray.  To illustrate how much this is the case, one is directed to the last entry into this saturated field, made by Lei et al (2018), on the centenary of M. Vaughan Williams'  birth. Their system is well-reasoned, comprehensive, inclusive of all existing agents, and therefore unwieldy and awkward:

• Class 0: Sinoatrial node blockers
  • Just one subclass, and one member - ivabradine
• Class I: voltage-gated sodium channel blockers
  • Class Ia: intermediate-dissociating Na⁺ channel blockers, eg. quinidine
  • Class Ib: fast-dissociating Na⁺ channel blockers, eg. lignocaine
  • Class Ic: slow-dissociating Na⁺ channel blockers, eg. flecainide
  • Class Id: late Na⁺ current blockers, eg. ranolazine
• Class II: Autonomic inhibitor and activators
  • Class IIa: non-selective β-blockers, eg. propanolol
  • Class IIb: non-selective β-agonists, eg. isoprenaline
  • Class IIc: Muscarinic M₂ receptor inhibitors, eg. atropine
  • Class IId: Muscarinic M₂ receptor agonists, eg. digoxin
  • Class IIe: Adenosine A₁ receptor agonists, eg. adenosine
• Class III: Potassium channel blockers and openers
  • Class IIIa: nonselective K⁺ channel blockers, eg. amiodarone and sotalol
  • Class IIIb: metabolically dependent K⁺ channel blockers, eg. nicorandil
  • Class IIIc: transmitter dependent K⁺ channel blockers (none available)
• Class IV: Ca²⁺ handling modulators
  • Class IVa: nonselective Ca²⁺ channel blockers (eg. verapamil and diltiazem)
  • Class IVb: intracellular Ca²⁺ channel blockers (eg. flecainide, propafenone)
  • Class IVc: sarcoplasmic reticular Ca²⁺ pump activators (none available)
  • Class IVd: membrane Ca²⁺ exchange inhibitors (none available)
  • Class IVe: Cytosolic Ca²⁺ handling protein phosphorylators (none)
• Class V: mechanosensitive channel blockers  (none currently available)
• Class VI: gap junction channel blockers (none curently available)
• Class VII: upstream target modulators  (ACE-inhibitors, statins, omega-3 fatty acids)

This system is reproduced here mainly because of the authors' fondness for awkward things, and in the hope that it one day finds more recognition than it currently has. One must remember that the official college textbook (and basically all the other influential material on this topic) were published before 2018 and will therefore be parroting the 1984 version of the Vaughan Williams classification. Moreover, many of the CICM Part I examiners will have trained during this time period and will have some nostalgic fondness for the old system. Therefore, the trainee is advised not to startle these people with any unnecessarily modern concepts. One should not present oneself as a dangerous radical anarchist during one's pharmacology viva. It is in this counterrevolutionary spirit that the rest of this discussion will be conducted.

Anyway.  From the emphasis on mechanisms of action which is inherent in this classification system, it follows that the only logical way to systematically discuss these agents is to start with their pharmacodynamic properties, and to handle the boring pharmacokinetics as an afterthought and shadow. This would be fairly consistent with college expectations, as they really only ever seemed interested in the ADME of amiodarone and digoxin, and those are handled separately.

Origins and mechanisms of arrhythmic and antiarrhythmic effects

To paraphrase the entire chapter dealing with abnormal cardiac electrical activity, as well as Grant (1992) and Barrio-Lopez (2020), arrhythmias arise because of:

• Abnormal automaticity, where some normal tissue becomes overexcited and decides to become a pacemaker, or existing pacemakers make pace in some disorganised or abnormal manner.
• Early afterdepolarisations, triggered depolarisations which occur during Phase 3, and which are promoted by anything which prolongs the repolarisation
• Late afterdepolarisations, triggered depolarisations which occur during Phase 4, and which are promoted by anything that might increase the intracellular calcium
• Reentry, where acton potential re-excites a patch of myocardium shortly after it has already depolarised, either because of some anatomical shortcut or because of an abnormally short refractory period

From this, it follows that antiarrhythmic effects should address these proarrhythmic mechanisms in some way. And indeed they do, but not in a way which makes them any easier to categorise. Many drugs with antiarrhythmic properties address several of these mechanisms simultaneously, and others may actually promote arrhythmias by prolonging repolarisation (those would be all the antiarrhythmics which prolong the QT interval). Still, it is possible to vaguely relate arrhythmogenic and antiarrhythmic mechanisms, as follows:

• Abnormal automaticity of normal pacemakers: it sounds like this would respond to anything that decreases the slope of Phase 4 in those cells, which would be calcium channel blockers and β-blockers
• Abnormal automaticity of non-pacemaker tissue should decrease as the result of beta-blockade. Reduced resting potential voltage of the membrane is usually the main problem, which these drugs should address.
• Late afterdepolarisations, where the problem is catecholamines or calcium, are addressed directly by calcium channel blockers and β-blockers
• Reentry, where action potential re-excites a patch of myocardium shortly after it has already depolarised, either because of some anatomical shortcut or because of an abnormally short refractory period. You can prolong the latter using Class Ia and Class III agents. As for abnormal conducting pathways, the best mechanism of managing this would be to slow conduction through those fibres, and this is where Class Ic agents should be ideal.

Now, for each class, some sort of short point-form summary of pharmacodynamic properties will be attempted, mainly in case one day these drugs become the topic of an SAQ. This has already happened to Class I agents in Question 9 from the second paper of 2012, which means that these are probably fair game. 

Electrophysiological effects of antiarrhythmic, in summary

In their comments to Question 9 from the second paper of 2012, the college had referred "an excellent table in Stoelting" as a way of describing what they wanted from a tabulated answer about the electrophysiological effects of Class 1 agents. That table is reproduced below, for easy reference: 

The reader must be warned that this table comes with no explanations in the text of that book, nor any references to follow. Still it remains a major reference for the people that write CICM exams. From this, we can conclude that it is intended to be committed to memory, and not to be understood. That would be enough to pass the CICM First Part exam. However, most normal people will probably agree that knowledge is not defined by data storage, and having a grasp of the underlying principles probably has some value for the intensivist in training. Therefore, wherever possible, some of the sections that follow will go on deep tangents into the electrophysiological trickery that produces these antiarrhythmic effects listed in the Stoelting table.

Class I antiarrhythmic agents: sodium channel blockers

Class I agents are sodium channel blockers. They generally bind to a site inside the pore of the Nav1.5 subunit of the fast voltage-gated sodium channel, which is responsible for Phase 0 of the cardiac action potential. All prefer to bind to open or inactivated sodium channels (though the slowly dissociating Class Ic agents remain bound even when the channels return to their resting state). Speaking of which, this class is further subdivided into subclasses according to what the drugs do to the action potential and what dissociation kinetics they have: 

• Class Ia agents, eg. quinidine and procainamide
  • Have an intermediate dissociation rate
  • Prolongs the duration of the action potential (mainly by their potassium channel blocker effects)
  • Therefore, prolong the QT interval
  • Prolong the QRS complex because of a longer Phase 0
  • Use-dependence: block effect (and QRS prolongation) is more pronounced in tachycardia because of slower dissociation from the binding site in diastole
• Class Ib agents, eg. lignocaine, mexelitine and phenytoin
  • Dissociate rapidly from the binding site, therefore free from use dependence
  • Have no effect on the duration of Phase 0
  • Therefore, do not prolong the QRS
  • Shorten the duration of the action potential, mainly by preventing late sustained sodium current
  • Therefore, shorten the QT interval
• Class Ic agents, eg. flecainide and propafenone
  • Dissociate slowly from the binding site, which means there is no use-dependence
  • Prolong Phase 0 more than other subclasses
  • Therefore, prolong the QRS duration
  • Have little effect on the duration of the action potential and therefore do not prolong the QT interval

The antiarrhythmic effect is felt in multiple ways:

• Automaticity of normal pacemakers should remain abnormal while you are on a Class I agent. Phase 4 of normal pacemaker cells does in fact depend on sodium currents (the "funny current", targeted by ivabradine), but those channels are distinct from the vast voltage-gated ones, and are not affected by Class I agents. Therefore, they generally don't tend to act as rhythm control agents for AF (for example). However, various reputable sources claim that they can still affect the slope of Phase 4 "by mechanisms not understood and unrelated to blocking fast sodium channels". Famously, flecainide has been used to suppress atrial fibrillation, and we still don't know how it exerts this effect (Echt & Ruskin, 2020). 
• Abnormal automaticity of non-pacemaker tissue should also remain unchanged, but it apparently decreases. It is also mentioned in various papers, but without much further explanation or reference. This is strange, as most people tend to think of Class I agents as anti-VT or anti-VF drugs.
• Early afterdepolarisations can actually increase in the context of Class I agent use, particularly with Class A agents which prolong the repolarisation. Like other QT-prolonging drugs, they can produce polymorphic VT. 
• Late afterdepolarisations should theoretically decrease in the context of sodium channel blockers, as they decrease the amount of intracellular sodium available for the sodium/calcium exchanger, which should theoretically mean less intracellular calcium being available (and it is the latter that is responsible for these phenomena). However, experimentally, this does not appear to be the case.
• Reentry is really where these agents are most effective. By decreasing the velocity of conduction, these agents slow the propagation of action potentials along the abnormal conducting pathways and therefore prevent reentrant tachycardias such as SVT. For this, Class Ic agents are ideal (for instance, flecainide is used to prevent arrhythmias in Wolff-Parkinson-White syndrome).

Class II antiarrhythmic agents: β-blockers

There is also a sub-classification of beta-blockers, one of which can be mentioned here:

• Non-selective
  • Propanolol
• β₁-selective
  • Atenolol
  • Metoprolol
  • Bisoprolol
  • Nebivolol
  • Esmolol
  • Sotalol
• Combined α- and β-blocker effect
  • Carvedilol
  • Labetalol

Though most of the antiarrhythmic effect comes from their β₁ effects, one needs to mention that some of these drugs have sodium channel blocker properties (propanolol) and others block potassium channels (sotalol).  However, they do not need those effects. Beta blockade has a rather diverse and potent effect on multiple cardiac ion channels, as listed in this table paraphrased from Dorian (2005):

Effect of Beta Blockers on Cardiac Ion Currents

Ion channel	Effect of beta-blockade
INa fast inward sodium current	Reduced current
Ito early, transient inward (repolarising) potassium current;	Reduced current
ICa,L L-type inward calcium current	Reduced current
INa/Ca  sodium/calcium exchange current	Reduced current
Iti  transient inward current	Reduced current
IK1 inward rectifier potassium current	Increased current
IKs  slow delayed rectifier potassium current	Reduced current
IKr  rapid delayed rectifier potassium current	Increased current
If  pacemaker current (sodium)	Reduced current

All of this is due to the downscaling of the intracellular cAMP-mediated signalling that results from beta blockers competing with catecholamines. According to the the table from Stoelting, the net electrophysiological effect of these ion channel current changes should be:

• an unchanged slope of Phase 0
• decreased conduction velocity
• decreased effective refractory period
• increased action potential duration
• decreased automaticity

which, on the surface ECG, should translate into

• a longer PR interval, and
• a shorter QT interval.

How does any of this connect to the abovementioned ion channel effects? For a drug class which has been available for so many decades, the literature on the electrophysiological effects of beta blockers is surprisingly well hidden. Some material can be unearthed by dusting off studies like Venditti et al (1987), who reported on the empirical findings of animal experiments for different beta blocker agents. The original table of results is presented below. As you can see, the situation is a lot more complex than the Stoelting table will have you believe:

The weird lack of data for commonly used drugs like metoprolol, and the weird excess of data for rare exotic beta blockers like pindolol and nadolol, is due to the vintage of this study, which dates back to 1987 (with metoprolol only having entered the market in 1982). Still, you can make out that broad trends really don't seem to exist across this group (for example, to say that all beta blockers uniformly increase the duration of the action potential would be patently untrue). Things are not helped by the fact that the best studied beta blocker was propanolol, as it was discovered way back in 1965 - but it also happens to have a bunch of sodium blocker effects which makes it a poor subject for studying pure β₁ antagonism. Things are definitely not helped by the fact that other studies produce totally contradictory fiundings, such as Sänchez-Chapula (1992) who looked at the effect of metoprolol on ventricular myocytes and found that it decreases the action potential duration.

Surely, by this point even the most loyal and patient reader will have sprayed profanities at their monitor, faced with this level of confusion and uncertainty. Who cares what happened to guinea pig myocytes in the eighties, they might ask. Tell me what CICM want me to write in my exam paper! Unfortunately, this absolutely reasonable request can only be answered with the Stoelting table. This is probably what the First Part Exam question-writers will be referring to when they put together the SAQs for the written paper. You can bet that they won't be leafing through yellowing copies of The American Journal of Cardiology from 1987. From this, it follows that Deranged Physiology would serve its readership best by trying to explain why the Stoelting table contains what it contains, and to find whatever evidence there is to support its assertions.  

Thus:

• An unchanged slope of Phase 0 should be a logical expectation of all beta blockers, so long as they can keep their hands off your sodium channels. The main determinant of this phase is the opening of the fast voltage-gated sodium channels which are generally supposed to be unaffected by beta blockade, as far as major textbooks seem to go. Propanolol, an agent with known sodium channel blocker effects, does seem to decrease the slope of this phase by about 30%, according to studies quoted in Venditti et al (1987), but the others supposedly do not. This seems to be contradicted by studies such as Murphy et al (1996) and Lu et al (1999) who found that these I_Na channels were markedly affected by beta-agonist stimulation (the total fast sodium current basically doubled when the myocytes were bathed in isoprenaline). This was not an effect on the function or properties of the channels themselves, but rather a result of a rather substantial increase in the number of available functional channels, which seemed to appear out of nowhere to help sodium cross the membrane. For Lu et al, this seemed to be a G-protein-related thing, and for Murphy et a, it was cAMP-mediated phosphorylation via PKA; in either case this means it should be susceptible to beta blockade. From this, it follows that beta blockers may actually reduce the slope of Phase 0, and therefore possibly even cause the QRS interval to widen (which the Stoelting table says they do).
• Decreased conduction velocity is a reported effect of beta blockers, mentioned in many leading textbooks and online resources. What this seems to be referring to is the conduction velocity through specialised conducting tissue, and the AV node specifically. Corino et al (2014) were able to demonstrate that the main reason for this is an increase in the effective refractory period of the AV node. In contrast, myocyte conduction velocity would probably be unaffected here. King et al (2013), in their excellent treatise on the determinants of cardiac myocyte conduction velocity, conclude that one of the most important factors is the rate of myocyte I_Na current (dV/dt), which is supposedly unaffected by beta blockers or beta agonists. If they did somehow affect those channels, beta blockers would also decrease the conduction velocity in cardiac muscle, but this does not seem to be a well established effect. When they discussed the mechanisms of beta blocker effectiveness in the treatment of serious ventricular arrhythmias, Reitter & Reiffel (1998) do not mention conduction velocity at all. 
• Decreased effective refractory period is listed in the Stoelting table as a property of Class II antiarrhythmics, but it does not sound like a favourable characteristic for an antiarrhythmic agent.  It is hard to find any evidence that this is correct, and in fact other textbooks and peer-reviewed publications in electrophysiology seem to confirm that beta blockers increase the ERP, and that a decreased ERP is actually a property of catecholamines like adrenaline.  Experimentally, Venditti et al (1987) reports that all beta blockers seem to prolong the refractory period of conductive tissues and myocytes, except for pindolol (which has an intrinsic direct sympathomimetic effect). In animal data, oxprenolol propanolol and timolol increased the ERP by 8%, 9% and 10%, respectively. In human data, looking specifically at something relevant to the intensivist (intravenous metoprolol), Camm et al (1982) could not find any significant signal in the effect on the atrial or ventricular refractory period (they increased in some patients, and decreased in others). In short, this may be a scenario where a textbook got something wrong. The CICM exam candidate may find themselves in the delicate position of having to consciously write an incorrect answer in order to score marks.
• Increased action potential duration is a logical extension of what we know about the effects of catecholamine β-agonists on cardiac action potential. From Dorian (2005), this slightly altered diagram of the beta-stimulated cardiac action potential demonstrates a reduction in the action potential duration:
  
  
  Thus, we should by all rights expect that beta blockers should produce the opposite effect. This seems to be supported by some ancient data from Edvardsson et al (1981), who demonstrated an average prolongation of the action potential by about at least 6%, and up to 22%, in some early hominid volunteers. The molecular mechanism for this seems to be calcium-related, as catecholamines exert their effects mainly by making intracellular calcium more available. These intracellular calcium concentration changes are a major determinant of total action potential duration. Specifically, high intracellular calcium concentrations result in a shorter action potential (Johansson & Wohlfart, 1980), mainly by favouring earlier repolarisation by their effect on the inward-rectifying potassium currents (i.e. probably Phase 2 is shortened, and Phase 3 begins sooner). From this, it follows that beta blockers should antagonise these phenomena by reducing the calcium entry.
• Decreased QTc duration is mentioned in the Stoelting table, and appears to be a real phenomenon, insofar as beta blockers are routinely used in the management of long QT syndrome and have a small but measurable effect on the QT interval which is rate-dependent (i.e. they prolong the QTc at slow heart rates and shorten it at faster heart rates).  We know that β-agonist catcholamines tend to prolong the QT interval, probably by delaying L-type calcium channel inactivation, and so it is logical to presume that beta blockers reverse this effect (though there does not seem to be any direct data in support of this statement).

So, how does all this affect arrhythmogenicity? In summary:

• Automaticity of normal pacemakers should decrease while under the effect of beta-blockers, as catecholamines are one of the main stimulants of this phenomenon. This is clearly demonstrated for most beta-blockers (Hernandez & Serrano, 1982). They do this probably by lowering the resting membrane potential during Phase 4, though it is not clear that this happens at conventional therapeutic concentrations.
• Abnormal automaticity of non-pacemaker tissue should also decrease, as beta-blockers are seen to reduce the appearance of ventricular ectopic beats. This is mainly because of the lower resting membrane potential.
• Early afterdepolarisations are reduced because the repolarisation time is reduced; this basically counteracts the proarrhythmic effect of QT-prolonging Class III agents, which is handy because many of them have pronounced beta blocker effects.
• Late afterdepolarisations, which are mainly related to intracellular calcium excess, tend to improve with anything that counteracts the effect of catecholamines (as they increase intracellular calcium concentrations).
• Reentry is often unaffected, as myocyte conduction velocity is largely unchanged in the presence of beta-blockade. The only exceptions may be specialised tissues such as the AV node, where unopposed vagal effects lead to conduction delay. Theoretically, that could defeat SVT by preventing the reentry of an abberantly conducted action potential.

Class III agents: potassium channel blockers

Amiodarone sotalol ibutilide and vernakalant are really the main contenders here, though worldwide there is an even larger selection of these agents available. This class prolongs repolarisation by interfering with the function of inward rectifier and outward delayed rectifier potassium currents, increasing the duration of the refractory period and of the action potential as a whole.

It is hard to discuss the Class III effects because the poster child for this class is amiodarone, and it acts promiscuously on all Class I-IV molecular targets. In fact, all Class III agents have some kind of extra weirdness (sotalol is a beta-blocker, ibutilide acts on slow inward depolarising sodium currents, etc). A "pure" potassium channel blocker effect is therefore difficult to describe using an example. Strictly speaking, they should only prolong repolarisation. The best resource to explain this would probably be the little fragment from Cardiac Electrophysiology: From Cell to Bedside (p. 518 of the 2017 edition). In short, these drugs mainly block Ikr, Iks and Ik1 currents which are responsible for Phase 3 of the cardiac action potential. Class III drugs are not unique in the effect, as there are many other drugs which interfere with this current (notably, macrolide antibiotics and antipsychotic drugs). One might say that some drugs prolong the QTc as an accidental side effect, but Class III agents do it intentionally. In the form of a diagram, one would express  this like so:

So, 

• Automaticity of normal pacemakers should remain more or less the same with a "pure" potassium blocker effect, but because amiodarone and sotalol have potent beta-blocker effects, these drugs are generally seen to slow the pacemaker rate.
• Abnormal automaticity of non-pacemaker tissue should also remain unchanged, but it probably increases, mainly because of early afterdepolarisations. Fortunately, most of these drugs have enough beta-blocking activity to counteract this.
• Early afterdepolarisations, which are mainly related to prolonged repolarisation, tend to increase because of a prolonged repolarisation time.
• Later afterdepolarisations due to a calcium-based mechanism are largely unaffected by Class III agents. 
• Reentry would be expected to decrease, as the prolonged effective refractory period should protect myocytes from adjacent ectopic pacemakers firing randomly.

Class IV agents: calcium channel blockers

Verapamil and diltiazem are the only real representatives here, as these are non-selective agents, whereas the dihydropyridine subclass tends to only affect the calcium channels in the vascular smooth muscle. Walker & Chia (1989) offer a good summary of their antiarrhythmic effects. In brief, their main effects are on pacemaker tissue, and on Phase 2 of the cardiac action potential. The Stoelting table makes a number of statements about these drugs, which need to be memorised unquestioningly by CICM exam candidates, and should not be subjected to any further scrutiny.

• No effect on conduction velocity, apparently, is the effect (lack of effect?) of calcium channel blockers. Looking at the literature, it is not clear where the authors of Stoelting got this from, as it is not supported by the experimental data for verapamil, nor reproduced in several other reputable resources. Perhaps the table may be referring purely to ventricular myocyte conduction, in which case it may be correct, as below the AV node the conduction velocity seems to be largely unaffected by therapeutic doses of verapamil (Bova et al, 1997). Clinically, the AV node seems to be the most important target of these agents, and slowing AV nodal conduction is probably how they seem to exert most of their antiarrhythmic effect (Wit & Cranefield, 1974).
  So: if the Stoelting table was referring purely to ventricular conduction velocity, then both beta blockers and calcium channel blockers should be listed as having no effect. If the Stoelting table was referring to AV nodal conduction velocity, then both groups should be listed as delaying it. Prystowsky (1988) is probably the best reference that compares the electrophysiology of these two drug classes, but at this stage, there does not appear to be any published experiment where the investigators laid out lengths of myocardium and raced impulses along them to contrast the effects.
• No effect on effective refractory period seems to be accurate, according to reputable sources, insofar as myocyte tissue is concerned (but the AV node effective refractory period is reliably prolonged). No effect on QT interval duration seems to be a logical upshot of this, as Phase 3 is unaffected. Both of these statements are contradicted by the finding that in long QT syndrome, verapamil shortens the QTc and decreases the effective refractory period of myocytes at high doses (Liu et al, 2016)
• Decreased action potential duration is listed in the Stoelting table, but this is not what the experimental evidence seems to suggest. Dersham & Han (1980) and Cranefield et al (1978) found that verapamil increased the APD in canine Purkinje fibres. On the other hand, FDA data sheets indicate that the FDA believes verapamil has no effect on action potential duration. Lastly, Yonekura et al (2018) found that verapamil decreased the APD in their weird fish model. 

In terms of their effects on arrhythmogenicity:

• Automaticity of normal pacemakers decreases with calcium channel antagonists, as they decrease the rate of Phase 0 rise in pacemaker cells (recall that their more gradual Phase 0 is purely due to L-type calcium channel opening). 
• Abnormal automaticity of non-pacemaker tissue should probably remain unchanged, as it is less reliant on calcium-mediated mechanisms. With this, calcium channel blockers should not be expected to control VT. The Stoelting table lists "no effect" under the automaticity effects of calcium channel blockers, which seems to suggest that they had meant "ectopic pacemaker automaticity"
• Early afterdepolarisations should be reduced by calcium channel blockers, as these are mainly related to prolonged repolarisation, and calcium channel blockers shorten repolarisation by decreasing the duration of Phase 2. 
• Later afterdepolarisations due to a calcium-based mechanism are the main course, and calcium channel blockers address their mechanism directly, by limiting the amount of available intracellular calcium
• Reentry would be expected to remain unaffected, except for the effect of calcium channel blockers on the performance of nodal tissue. AV nodal and atrial bundle tissue conduction should decrease, and this could put an end to SVT. 

Other and Misc

At risk of being seen to promote the apostate classification system by Lei et al (2018), one needs to tentatively recognise that there are other drugs out there (digoxin, magnesium, adenosine, etc) which do not fit neatly into the Vaughan Williams classification. 

• Digoxin and the other cardiac glycosides mainly exert their antiarrhythmic effects by a vagomimetic action on the AV node, where they slow the conduction of action potentials. By increasing the availability of intracellular calcium, digoxin may lead to an increased risk of late afterdepolarisations. Other than that, it has no real positive effect on the arrhythmogenesis of ventricular myocardium (Gbadebo et al, 2000)
• Adenosine can be described as an antiarrhythmic by virtue of the fact that we only tend to use it for patients with arrhythmias, specifically with SVT. It also has some anti-adrenergic functions (think of it doing the opposite of everything coffee does), which can potentially have a beta-blocker like effect on ventricular arrhythmogenesis (Wilbur et al, 1997)
• Magnesium has antiarrhythmic properties which are most pronounced in those arrhythmias where the mechanism involves early or delayed afterdepolarisations. It can be classified along with Class III and Class IV agents, as its effects are mainly involved in decreasing the duration of repolarisation by acting as an antagonist of potassium and calcium currents (Fazekas et al, 1993). By the same effect, it counteracts the depolarisation of pacemaker tissues, like a calcium channel blocker.